yt_dlp/extractor/cbc.py

   1 import re
   2 import json
   3 import base64
   4 import time
   5 import urllib.parse
   6
   7 from .common import InfoExtractor
   8 from ..compat import (
   9     compat_str,
  10 )
  11 from ..utils import (
  12     ExtractorError,
  13     int_or_none,
  14     join_nonempty,
  15     js_to_json,
  16     orderedSet,
  17     parse_iso8601,
  18     smuggle_url,
  19     strip_or_none,
  20     traverse_obj,
  21     try_get,
  22 )
  23
  24
  25 class CBCIE(InfoExtractor):
  26     IE_NAME = 'cbc.ca'
  27     _VALID_URL = r'https?://(?:www\.)?cbc\.ca/(?!player/)(?:[^/]+/)+(?P<id>[^/?#]+)'
  28     _TESTS = [{
  29         # with mediaId
  30         'url': 'http://www.cbc.ca/22minutes/videos/clips-season-23/don-cherry-play-offs',
  31         'md5': '97e24d09672fc4cf56256d6faa6c25bc',
  32         'info_dict': {
  33             'id': '2682904050',
  34             'ext': 'mp4',
  35             'title': 'Don Cherry – All-Stars',
  36             'description': 'Don Cherry has a bee in his bonnet about AHL player John Scott because that guy’s got heart.',
  37             'timestamp': 1454463000,
  38             'upload_date': '20160203',
  39             'uploader': 'CBCC-NEW',
  40         },
  41         'skip': 'Geo-restricted to Canada',
  42     }, {
  43         # with clipId, feed available via tpfeed.cbc.ca and feed.theplatform.com
  44         'url': 'http://www.cbc.ca/22minutes/videos/22-minutes-update/22-minutes-update-episode-4',
  45         'md5': '162adfa070274b144f4fdc3c3b8207db',
  46         'info_dict': {
  47             'id': '2414435309',
  48             'ext': 'mp4',
  49             'title': '22 Minutes Update: What Not To Wear Quebec',
  50             'description': "This week's latest Canadian top political story is What Not To Wear Quebec.",
  51             'upload_date': '20131025',
  52             'uploader': 'CBCC-NEW',
  53             'timestamp': 1382717907,
  54         },
  55         'skip': 'No longer available',
  56     }, {
  57         # with clipId, feed only available via tpfeed.cbc.ca
  58         'url': 'http://www.cbc.ca/archives/entry/1978-robin-williams-freestyles-on-90-minutes-live',
  59         'md5': '0274a90b51a9b4971fe005c63f592f12',
  60         'info_dict': {
  61             'id': '2487345465',
  62             'ext': 'mp4',
  63             'title': 'Robin Williams freestyles on 90 Minutes Live',
  64             'description': 'Wacky American comedian Robin Williams shows off his infamous "freestyle" comedic talents while being interviewed on CBC\'s 90 Minutes Live.',
  65             'upload_date': '19780210',
  66             'uploader': 'CBCC-NEW',
  67             'timestamp': 255977160,
  68         },
  69         'skip': '404 Not Found',
  70     }, {
  71         # multiple iframes
  72         'url': 'http://www.cbc.ca/natureofthings/blog/birds-eye-view-from-vancouvers-burrard-street-bridge-how-we-got-the-shot',
  73         'playlist': [{
  74             'md5': '377572d0b49c4ce0c9ad77470e0b96b4',
  75             'info_dict': {
  76                 'id': '2680832926',
  77                 'ext': 'mp4',
  78                 'title': 'An Eagle\'s-Eye View Off Burrard Bridge',
  79                 'description': 'Hercules the eagle flies from Vancouver\'s Burrard Bridge down to a nearby park with a mini-camera strapped to his back.',
  80                 'upload_date': '20160201',
  81                 'timestamp': 1454342820,
  82                 'uploader': 'CBCC-NEW',
  83             },
  84         }, {
  85             'md5': '415a0e3f586113894174dfb31aa5bb1a',
  86             'info_dict': {
  87                 'id': '2658915080',
  88                 'ext': 'mp4',
  89                 'title': 'Fly like an eagle!',
  90                 'description': 'Eagle equipped with a mini camera flies from the world\'s tallest tower',
  91                 'upload_date': '20150315',
  92                 'timestamp': 1426443984,
  93                 'uploader': 'CBCC-NEW',
  94             },
  95         }],
  96         'skip': 'Geo-restricted to Canada',
  97     }, {
  98         # multiple CBC.APP.Caffeine.initInstance(...)
  99         'url': 'http://www.cbc.ca/news/canada/calgary/dog-indoor-exercise-winter-1.3928238',
 100         'info_dict': {
 101             'title': 'Keep Rover active during the deep freeze with doggie pushups and other fun indoor tasks',  # FIXME
 102             'id': 'dog-indoor-exercise-winter-1.3928238',
 103             'description': 'md5:c18552e41726ee95bd75210d1ca9194c',
 104         },
 105         'playlist_mincount': 6,
 106     }]
 107
 108     @classmethod
 109     def suitable(cls, url):
 110         return False if CBCPlayerIE.suitable(url) else super(CBCIE, cls).suitable(url)
 111
 112     def _extract_player_init(self, player_init, display_id):
 113         player_info = self._parse_json(player_init, display_id, js_to_json)
 114         media_id = player_info.get('mediaId')
 115         if not media_id:
 116             clip_id = player_info['clipId']
 117             feed = self._download_json(
 118                 'http://tpfeed.cbc.ca/f/ExhSPC/vms_5akSXx4Ng_Zn?byCustomValue={:mpsReleases}{%s}' % clip_id,
 119                 clip_id, fatal=False)
 120             if feed:
 121                 media_id = try_get(feed, lambda x: x['entries'][0]['guid'], compat_str)
 122             if not media_id:
 123                 media_id = self._download_json(
 124                     'http://feed.theplatform.com/f/h9dtGB/punlNGjMlc1F?fields=id&byContent=byReleases%3DbyId%253D' + clip_id,
 125                     clip_id)['entries'][0]['id'].split('/')[-1]
 126         return self.url_result('cbcplayer:%s' % media_id, 'CBCPlayer', media_id)
 127
 128     def _real_extract(self, url):
 129         display_id = self._match_id(url)
 130         webpage = self._download_webpage(url, display_id)
 131         title = (self._og_search_title(webpage, default=None)
 132                  or self._html_search_meta('twitter:title', webpage, 'title', default=None)
 133                  or self._html_extract_title(webpage))
 134         entries = [
 135             self._extract_player_init(player_init, display_id)
 136             for player_init in re.findall(r'CBC\.APP\.Caffeine\.initInstance\(({.+?})\);', webpage)]
 137         media_ids = []
 138         for media_id_re in (
 139                 r'<iframe[^>]+src="[^"]+?mediaId=(\d+)"',
 140                 r'<div[^>]+\bid=["\']player-(\d+)',
 141                 r'guid["\']\s*:\s*["\'](\d+)'):
 142             media_ids.extend(re.findall(media_id_re, webpage))
 143         entries.extend([
 144             self.url_result('cbcplayer:%s' % media_id, 'CBCPlayer', media_id)
 145             for media_id in orderedSet(media_ids)])
 146         return self.playlist_result(
 147             entries, display_id, strip_or_none(title),
 148             self._og_search_description(webpage))
 149
 150
 151 class CBCPlayerIE(InfoExtractor):
 152     IE_NAME = 'cbc.ca:player'
 153     _VALID_URL = r'(?:cbcplayer:|https?://(?:www\.)?cbc\.ca/(?:player/play/|i/caffeine/syndicate/\?mediaId=))(?P<id>\d+)'
 154     _TESTS = [{
 155         'url': 'http://www.cbc.ca/player/play/2683190193',
 156         'md5': '64d25f841ddf4ddb28a235338af32e2c',
 157         'info_dict': {
 158             'id': '2683190193',
 159             'ext': 'mp4',
 160             'title': 'Gerry Runs a Sweat Shop',
 161             'description': 'md5:b457e1c01e8ff408d9d801c1c2cd29b0',
 162             'timestamp': 1455071400,
 163             'upload_date': '20160210',
 164             'uploader': 'CBCC-NEW',
 165         },
 166         'skip': 'Geo-restricted to Canada and no longer available',
 167     }, {
 168         # Redirected from http://www.cbc.ca/player/AudioMobile/All%20in%20a%20Weekend%20Montreal/ID/2657632011/
 169         'url': 'http://www.cbc.ca/player/play/2657631896',
 170         'md5': 'e5e708c34ae6fca156aafe17c43e8b75',
 171         'info_dict': {
 172             'id': '2657631896',
 173             'ext': 'mp3',
 174             'title': 'CBC Montreal is organizing its first ever community hackathon!',
 175             'description': 'The modern technology we tend to depend on so heavily, is never without it\'s share of hiccups and headaches. Next weekend - CBC Montreal will be getting members of the public for its first Hackathon.',
 176             'timestamp': 1425704400,
 177             'upload_date': '20150307',
 178             'uploader': 'CBCC-NEW',
 179             'thumbnail': 'http://thumbnails.cbc.ca/maven_legacy/thumbnails/sonali-karnick-220.jpg',
 180             'chapters': [],
 181             'duration': 494.811,
 182         },
 183     }, {
 184         'url': 'http://www.cbc.ca/player/play/2164402062',
 185         'md5': '33fcd8f6719b9dd60a5e73adcb83b9f6',
 186         'info_dict': {
 187             'id': '2164402062',
 188             'ext': 'mp4',
 189             'title': 'Cancer survivor four times over',
 190             'description': 'Tim Mayer has beaten three different forms of cancer four times in five years.',
 191             'timestamp': 1320410746,
 192             'upload_date': '20111104',
 193             'uploader': 'CBCC-NEW',
 194             'thumbnail': 'https://thumbnails.cbc.ca/maven_legacy/thumbnails/277/67/cancer_852x480_2164412612.jpg',
 195             'chapters': [],
 196             'duration': 186.867,
 197         },
 198     }, {
 199         # Has subtitles
 200         # These broadcasts expire after ~1 month, can find new test URL here:
 201         # https://www.cbc.ca/player/news/TV%20Shows/The%20National/Latest%20Broadcast
 202         'url': 'http://www.cbc.ca/player/play/2249992771553',
 203         'md5': '2f2fb675dd4f0f8a5bb7588d1b13bacd',
 204         'info_dict': {
 205             'id': '2249992771553',
 206             'ext': 'mp4',
 207             'title': 'The National | Women’s soccer pay, Florida seawater, Swift quake',
 208             'description': 'md5:adba28011a56cfa47a080ff198dad27a',
 209             'timestamp': 1690596000,
 210             'duration': 2716.333,
 211             'subtitles': {'eng': [{'ext': 'vtt', 'protocol': 'm3u8_native'}]},
 212             'thumbnail': 'https://thumbnails.cbc.ca/maven_legacy/thumbnails/481/326/thumbnail.jpeg',
 213             'uploader': 'CBCC-NEW',
 214             'chapters': 'count:5',
 215             'upload_date': '20230729',
 216         },
 217     }]
 218
 219     def _real_extract(self, url):
 220         video_id = self._match_id(url)
 221         return {
 222             '_type': 'url_transparent',
 223             'ie_key': 'ThePlatform',
 224             'url': smuggle_url(
 225                 'http://link.theplatform.com/s/ExhSPC/media/guid/2655402169/%s?mbr=true&formats=MPEG4,FLV,MP3' % video_id, {
 226                     'force_smil_url': True
 227                 }),
 228             'id': video_id,
 229             '_format_sort_fields': ('res', 'proto')  # Prioritize direct http formats over HLS
 230         }
 231
 232
 233 class CBCPlayerPlaylistIE(InfoExtractor):
 234     IE_NAME = 'cbc.ca:player:playlist'
 235     _VALID_URL = r'https?://(?:www\.)?cbc\.ca/(?:player/)(?!play/)(?P<id>[^?#]+)'
 236     _TESTS = [{
 237         'url': 'https://www.cbc.ca/player/news/TV%20Shows/The%20National/Latest%20Broadcast',
 238         'playlist_mincount': 25,
 239         'info_dict': {
 240             'id': 'news/tv shows/the national/latest broadcast',
 241         }
 242     }, {
 243         'url': 'https://www.cbc.ca/player/news/Canada/North',
 244         'playlist_mincount': 25,
 245         'info_dict': {
 246             'id': 'news/canada/north',
 247         }
 248     }]
 249
 250     def _real_extract(self, url):
 251         playlist_id = urllib.parse.unquote(self._match_id(url)).lower()
 252         webpage = self._download_webpage(url, playlist_id)
 253         json_content = self._search_json(
 254             r'window\.__INITIAL_STATE__\s*=', webpage, 'initial state', playlist_id)
 255
 256         def entries():
 257             for video_id in traverse_obj(json_content, (
 258                 'video', 'clipsByCategory', lambda k, _: k.lower() == playlist_id, 'items', ..., 'id'
 259             )):
 260                 yield self.url_result(f'https://www.cbc.ca/player/play/{video_id}', CBCPlayerIE)
 261
 262         return self.playlist_result(entries(), playlist_id)
 263
 264
 265 class CBCGemIE(InfoExtractor):
 266     IE_NAME = 'gem.cbc.ca'
 267     _VALID_URL = r'https?://gem\.cbc\.ca/(?:media/)?(?P<id>[0-9a-z-]+/s[0-9]+[a-z][0-9]+)'
 268     _TESTS = [{
 269         # This is a normal, public, TV show video
 270         'url': 'https://gem.cbc.ca/media/schitts-creek/s06e01',
 271         'md5': '93dbb31c74a8e45b378cf13bd3f6f11e',
 272         'info_dict': {
 273             'id': 'schitts-creek/s06e01',
 274             'ext': 'mp4',
 275             'title': 'Smoke Signals',
 276             'description': 'md5:929868d20021c924020641769eb3e7f1',
 277             'thumbnail': 'https://images.radio-canada.ca/v1/synps-cbc/episode/perso/cbc_schitts_creek_season_06e01_thumbnail_v01.jpg?im=Resize=(Size)',
 278             'duration': 1314,
 279             'categories': ['comedy'],
 280             'series': 'Schitt\'s Creek',
 281             'season': 'Season 6',
 282             'season_number': 6,
 283             'episode': 'Smoke Signals',
 284             'episode_number': 1,
 285             'episode_id': 'schitts-creek/s06e01',
 286         },
 287         'params': {'format': 'bv'},
 288         'skip': 'Geo-restricted to Canada',
 289     }, {
 290         # This video requires an account in the browser, but works fine in yt-dlp
 291         'url': 'https://gem.cbc.ca/media/schitts-creek/s01e01',
 292         'md5': '297a9600f554f2258aed01514226a697',
 293         'info_dict': {
 294             'id': 'schitts-creek/s01e01',
 295             'ext': 'mp4',
 296             'title': 'The Cup Runneth Over',
 297             'description': 'md5:9bca14ea49ab808097530eb05a29e797',
 298             'thumbnail': 'https://images.radio-canada.ca/v1/synps-cbc/episode/perso/cbc_schitts_creek_season_01e01_thumbnail_v01.jpg?im=Resize=(Size)',
 299             'series': 'Schitt\'s Creek',
 300             'season_number': 1,
 301             'season': 'Season 1',
 302             'episode_number': 1,
 303             'episode': 'The Cup Runneth Over',
 304             'episode_id': 'schitts-creek/s01e01',
 305             'duration': 1309,
 306             'categories': ['comedy'],
 307         },
 308         'params': {'format': 'bv'},
 309         'skip': 'Geo-restricted to Canada',
 310     }, {
 311         'url': 'https://gem.cbc.ca/nadiyas-family-favourites/s01e01',
 312         'only_matching': True,
 313     }]
 314
 315     _GEO_COUNTRIES = ['CA']
 316     _TOKEN_API_KEY = '3f4beddd-2061-49b0-ae80-6f1f2ed65b37'
 317     _NETRC_MACHINE = 'cbcgem'
 318     _claims_token = None
 319
 320     def _new_claims_token(self, email, password):
 321         data = json.dumps({
 322             'email': email,
 323             'password': password,
 324         }).encode()
 325         headers = {'content-type': 'application/json'}
 326         query = {'apikey': self._TOKEN_API_KEY}
 327         resp = self._download_json('https://api.loginradius.com/identity/v2/auth/login',
 328                                    None, data=data, headers=headers, query=query)
 329         access_token = resp['access_token']
 330
 331         query = {
 332             'access_token': access_token,
 333             'apikey': self._TOKEN_API_KEY,
 334             'jwtapp': 'jwt',
 335         }
 336         resp = self._download_json('https://cloud-api.loginradius.com/sso/jwt/api/token',
 337                                    None, headers=headers, query=query)
 338         sig = resp['signature']
 339
 340         data = json.dumps({'jwt': sig}).encode()
 341         headers = {'content-type': 'application/json', 'ott-device-type': 'web'}
 342         resp = self._download_json('https://services.radio-canada.ca/ott/cbc-api/v2/token',
 343                                    None, data=data, headers=headers, expected_status=426)
 344         cbc_access_token = resp['accessToken']
 345
 346         headers = {'content-type': 'application/json', 'ott-device-type': 'web', 'ott-access-token': cbc_access_token}
 347         resp = self._download_json('https://services.radio-canada.ca/ott/cbc-api/v2/profile',
 348                                    None, headers=headers, expected_status=426)
 349         return resp['claimsToken']
 350
 351     def _get_claims_token_expiry(self):
 352         # Token is a JWT
 353         # JWT is decoded here and 'exp' field is extracted
 354         # It is a Unix timestamp for when the token expires
 355         b64_data = self._claims_token.split('.')[1]
 356         data = base64.urlsafe_b64decode(b64_data + "==")
 357         return json.loads(data)['exp']
 358
 359     def claims_token_expired(self):
 360         exp = self._get_claims_token_expiry()
 361         if exp - time.time() < 10:
 362             # It will expire in less than 10 seconds, or has already expired
 363             return True
 364         return False
 365
 366     def claims_token_valid(self):
 367         return self._claims_token is not None and not self.claims_token_expired()
 368
 369     def _get_claims_token(self, email, password):
 370         if not self.claims_token_valid():
 371             self._claims_token = self._new_claims_token(email, password)
 372             self.cache.store(self._NETRC_MACHINE, 'claims_token', self._claims_token)
 373         return self._claims_token
 374
 375     def _real_initialize(self):
 376         if self.claims_token_valid():
 377             return
 378         self._claims_token = self.cache.load(self._NETRC_MACHINE, 'claims_token')
 379
 380     def _find_secret_formats(self, formats, video_id):
 381         """ Find a valid video url and convert it to the secret variant """
 382         base_format = next((f for f in formats if f.get('vcodec') != 'none'), None)
 383         if not base_format:
 384             return
 385
 386         base_url = re.sub(r'(Manifest\(.*?),filter=[\w-]+(.*?\))', r'\1\2', base_format['url'])
 387         url = re.sub(r'(Manifest\(.*?),format=[\w-]+(.*?\))', r'\1\2', base_url)
 388
 389         secret_xml = self._download_xml(url, video_id, note='Downloading secret XML', fatal=False)
 390         if not secret_xml:
 391             return
 392
 393         for child in secret_xml:
 394             if child.attrib.get('Type') != 'video':
 395                 continue
 396             for video_quality in child:
 397                 bitrate = int_or_none(video_quality.attrib.get('Bitrate'))
 398                 if not bitrate or 'Index' not in video_quality.attrib:
 399                     continue
 400                 height = int_or_none(video_quality.attrib.get('MaxHeight'))
 401
 402                 yield {
 403                     **base_format,
 404                     'format_id': join_nonempty('sec', height),
 405                     # Note: \g<1> is necessary instead of \1 since bitrate is a number
 406                     'url': re.sub(r'(QualityLevels\()\d+(\))', fr'\g<1>{bitrate}\2', base_url),
 407                     'width': int_or_none(video_quality.attrib.get('MaxWidth')),
 408                     'tbr': bitrate / 1000.0,
 409                     'height': height,
 410                 }
 411
 412     def _real_extract(self, url):
 413         video_id = self._match_id(url)
 414         video_info = self._download_json(
 415             f'https://services.radio-canada.ca/ott/cbc-api/v2/assets/{video_id}',
 416             video_id, expected_status=426)
 417
 418         email, password = self._get_login_info()
 419         if email and password:
 420             claims_token = self._get_claims_token(email, password)
 421             headers = {'x-claims-token': claims_token}
 422         else:
 423             headers = {}
 424         m3u8_info = self._download_json(video_info['playSession']['url'], video_id, headers=headers)
 425         m3u8_url = m3u8_info.get('url')
 426
 427         if m3u8_info.get('errorCode') == 1:
 428             self.raise_geo_restricted(countries=['CA'])
 429         elif m3u8_info.get('errorCode') == 35:
 430             self.raise_login_required(method='password')
 431         elif m3u8_info.get('errorCode') != 0:
 432             raise ExtractorError(f'{self.IE_NAME} said: {m3u8_info.get("errorCode")} - {m3u8_info.get("message")}')
 433
 434         formats = self._extract_m3u8_formats(m3u8_url, video_id, m3u8_id='hls')
 435         self._remove_duplicate_formats(formats)
 436         formats.extend(self._find_secret_formats(formats, video_id))
 437
 438         for format in formats:
 439             if format.get('vcodec') == 'none':
 440                 if format.get('ext') is None:
 441                     format['ext'] = 'm4a'
 442                 if format.get('acodec') is None:
 443                     format['acodec'] = 'mp4a.40.2'
 444
 445                 # Put described audio at the beginning of the list, so that it
 446                 # isn't chosen by default, as most people won't want it.
 447                 if 'descriptive' in format['format_id'].lower():
 448                     format['preference'] = -2
 449
 450         return {
 451             'id': video_id,
 452             'title': video_info['title'],
 453             'description': video_info.get('description'),
 454             'thumbnail': video_info.get('image'),
 455             'series': video_info.get('series'),
 456             'season_number': video_info.get('season'),
 457             'season': f'Season {video_info.get("season")}',
 458             'episode_number': video_info.get('episode'),
 459             'episode': video_info.get('title'),
 460             'episode_id': video_id,
 461             'duration': video_info.get('duration'),
 462             'categories': [video_info.get('category')],
 463             'formats': formats,
 464             'release_timestamp': video_info.get('airDate'),
 465             'timestamp': video_info.get('availableDate'),
 466         }
 467
 468
 469 class CBCGemPlaylistIE(InfoExtractor):
 470     IE_NAME = 'gem.cbc.ca:playlist'
 471     _VALID_URL = r'https?://gem\.cbc\.ca/(?:media/)?(?P<id>(?P<show>[0-9a-z-]+)/s(?P<season>[0-9]+))/?(?:[?#]|$)'
 472     _TESTS = [{
 473         # TV show playlist, all public videos
 474         'url': 'https://gem.cbc.ca/media/schitts-creek/s06',
 475         'playlist_count': 16,
 476         'info_dict': {
 477             'id': 'schitts-creek/s06',
 478             'title': 'Season 6',
 479             'description': 'md5:6a92104a56cbeb5818cc47884d4326a2',
 480             'series': 'Schitt\'s Creek',
 481             'season_number': 6,
 482             'season': 'Season 6',
 483             'thumbnail': 'https://images.radio-canada.ca/v1/synps-cbc/season/perso/cbc_schitts_creek_season_06_carousel_v03.jpg?impolicy=ott&im=Resize=(_Size_)&quality=75',
 484         },
 485     }, {
 486         'url': 'https://gem.cbc.ca/schitts-creek/s06',
 487         'only_matching': True,
 488     }]
 489     _API_BASE = 'https://services.radio-canada.ca/ott/cbc-api/v2/shows/'
 490
 491     def _real_extract(self, url):
 492         match = self._match_valid_url(url)
 493         season_id = match.group('id')
 494         show = match.group('show')
 495         show_info = self._download_json(self._API_BASE + show, season_id, expected_status=426)
 496         season = int(match.group('season'))
 497
 498         season_info = next((s for s in show_info['seasons'] if s.get('season') == season), None)
 499
 500         if season_info is None:
 501             raise ExtractorError(f'Couldn\'t find season {season} of {show}')
 502
 503         episodes = []
 504         for episode in season_info['assets']:
 505             episodes.append({
 506                 '_type': 'url_transparent',
 507                 'ie_key': 'CBCGem',
 508                 'url': 'https://gem.cbc.ca/media/' + episode['id'],
 509                 'id': episode['id'],
 510                 'title': episode.get('title'),
 511                 'description': episode.get('description'),
 512                 'thumbnail': episode.get('image'),
 513                 'series': episode.get('series'),
 514                 'season_number': episode.get('season'),
 515                 'season': season_info['title'],
 516                 'season_id': season_info.get('id'),
 517                 'episode_number': episode.get('episode'),
 518                 'episode': episode.get('title'),
 519                 'episode_id': episode['id'],
 520                 'duration': episode.get('duration'),
 521                 'categories': [episode.get('category')],
 522             })
 523
 524         thumbnail = None
 525         tn_uri = season_info.get('image')
 526         # the-national was observed to use a "data:image/png;base64"
 527         # URI for their 'image' value. The image was 1x1, and is
 528         # probably just a placeholder, so it is ignored.
 529         if tn_uri is not None and not tn_uri.startswith('data:'):
 530             thumbnail = tn_uri
 531
 532         return {
 533             '_type': 'playlist',
 534             'entries': episodes,
 535             'id': season_id,
 536             'title': season_info['title'],
 537             'description': season_info.get('description'),
 538             'thumbnail': thumbnail,
 539             'series': show_info.get('title'),
 540             'season_number': season_info.get('season'),
 541             'season': season_info['title'],
 542         }
 543
 544
 545 class CBCGemLiveIE(InfoExtractor):
 546     IE_NAME = 'gem.cbc.ca:live'
 547     _VALID_URL = r'https?://gem\.cbc\.ca/live(?:-event)?/(?P<id>\d+)'
 548     _TESTS = [
 549         {
 550             'url': 'https://gem.cbc.ca/live/920604739687',
 551             'info_dict': {
 552                 'title': 'Ottawa',
 553                 'description': 'The live TV channel and local programming from Ottawa',
 554                 'thumbnail': 'https://thumbnails.cbc.ca/maven_legacy/thumbnails/CBC_OTT_VMS/Live_Channel_Static_Images/Ottawa_2880x1620.jpg',
 555                 'is_live': True,
 556                 'id': 'AyqZwxRqh8EH',
 557                 'ext': 'mp4',
 558                 'timestamp': 1492106160,
 559                 'upload_date': '20170413',
 560                 'uploader': 'CBCC-NEW',
 561             },
 562             'skip': 'Live might have ended',
 563         },
 564         {
 565             'url': 'https://gem.cbc.ca/live/44',
 566             'info_dict': {
 567                 'id': '44',
 568                 'ext': 'mp4',
 569                 'is_live': True,
 570                 'title': r're:^Ottawa [0-9\-: ]+',
 571                 'description': 'The live TV channel and local programming from Ottawa',
 572                 'live_status': 'is_live',
 573                 'thumbnail': r're:https://images.gem.cbc.ca/v1/cbc-gem/live/.*'
 574             },
 575             'params': {'skip_download': True},
 576             'skip': 'Live might have ended',
 577         },
 578         {
 579             'url': 'https://gem.cbc.ca/live-event/10835',
 580             'info_dict': {
 581                 'id': '10835',
 582                 'ext': 'mp4',
 583                 'is_live': True,
 584                 'title': r're:^The National \| Biden’s trip wraps up, Paltrow testifies, Bird flu [0-9\-: ]+',
 585                 'description': 'March 24, 2023 | President Biden’s Ottawa visit ends with big pledges from both countries. Plus, Gwyneth Paltrow testifies in her ski collision trial.',
 586                 'live_status': 'is_live',
 587                 'thumbnail': r're:https://images.gem.cbc.ca/v1/cbc-gem/live/.*',
 588                 'timestamp': 1679706000,
 589                 'upload_date': '20230325',
 590             },
 591             'params': {'skip_download': True},
 592             'skip': 'Live might have ended',
 593         }
 594     ]
 595
 596     def _real_extract(self, url):
 597         video_id = self._match_id(url)
 598         webpage = self._download_webpage(url, video_id)
 599         video_info = self._search_nextjs_data(webpage, video_id)['props']['pageProps']['data']
 600
 601         # Two types of metadata JSON
 602         if not video_info.get('formattedIdMedia'):
 603             video_info = traverse_obj(
 604                 video_info, (('freeTv', ('streams', ...)), 'items', lambda _, v: v['key'] == video_id, {dict}),
 605                 get_all=False, default={})
 606
 607         video_stream_id = video_info.get('formattedIdMedia')
 608         if not video_stream_id:
 609             raise ExtractorError('Couldn\'t find video metadata, maybe this livestream is now offline', expected=True)
 610
 611         stream_data = self._download_json(
 612             'https://services.radio-canada.ca/media/validation/v2/', video_id, query={
 613                 'appCode': 'mpx',
 614                 'connectionType': 'hd',
 615                 'deviceType': 'ipad',
 616                 'idMedia': video_stream_id,
 617                 'multibitrate': 'true',
 618                 'output': 'json',
 619                 'tech': 'hls',
 620                 'manifestType': 'desktop',
 621             })
 622
 623         return {
 624             'id': video_id,
 625             'formats': self._extract_m3u8_formats(stream_data['url'], video_id, 'mp4', live=True),
 626             'is_live': True,
 627             **traverse_obj(video_info, {
 628                 'title': 'title',
 629                 'description': 'description',
 630                 'thumbnail': ('images', 'card', 'url'),
 631                 'timestamp': ('airDate', {parse_iso8601}),
 632             })
 633         }